Millions of diverse, older adults and their surrogate decision makers will face many complex, ongoing decisions over the course of chronic and advanced illness. Preparation for decision is making or advance care planning (ACP), has traditionally focused on making decisions about life-prolonging procedures (e.g., mechanical ventilation) by completing advance directives. Yet, advance directives often fail to prepare patients to identify their values or to communicate with surrogates and clinicians. Therefore, the research team developed a broader paradigm of ACP that focuses instead on preparing patients to communicate their wishes and to participate with clinicians in making real-time, complex medical decisions over the course of illness. The research team has created an easy-to-understand, patient-centered website called PREPARE that is based on this new ACP paradigm. PREPARE teaches patients, through videos and modeling of ACP behaviors, how to identify their wishes and communicate with surrogates and clinicians. The objective of this proposal is to test PREPARE in a double-blind, randomized, efficacy trial. The aims of this study are: (1) to determine the efficacy of PREPARE to engage ethnically diverse, older adults in a range of ACP behaviors (i.e., identify and discuss wishes with surrogates and clinicians as well as complete advance directives), and (2) to assess whether the efficacy of PREPARE varies by patient subgroups, such as race/ethnicity and literacy level. The researchers will also obtain input from patients, surrogates, and clinicians about implementation of PREPARE in safety-net settings. Methods: 500 patients aged e 55 years with chronic illness from San Francisco General Hospital will be randomly assigned to either the intervention arm (PREPARE plus an advance directive) or the control arm (advance directive only). Main outcomes of patient- reported engagement in ACP will be measured with validated surveys at 1 week, 3, 6, and 12 months (Aim 1). We will use mixed-effects generalized linear models to assess differences in longitudinal outcomes between arms and control for variables that differ between arms. We will also adjust for clustering by clinician. To assess efficacy by patient subgroups, we will include interaction terms of race/ethnicity and literacy by study arm in the modeling (Aim 2). Relevance to NIH and public health: If PREPARE is efficacious, the researchers will have a novel, practical, scalable ACP guide that will be easy to use and disseminate within diverse populations. This will result in diverse, older adults who are prepared to communicate their wishes and make complex medical decisions over the course of serious and chronic illness.